This section compiles results from developing and pilot-testing a mobile augmented reality tool, where more detailed information can be found in Appendix 2, 3 and 6. It first describes the specifications and design criteria developed for the designing of the tool. Then the section describes the implementation of the design criteria during six iterative prototyping sessions. The following section accounts for the assessments and takeaways from these six prototypes and describes the two resulting semi-pilot systems of the Urban CoBuilder: a 1:1 scale outdoor mobile augmented reality (MAR) version, and a scalable table-top version augmented reality (AR) tool. Subsequently, an assessment of these two semi-pilot systems is provided.
5.3.3.1 Specifications and design criteria
The takeaways from the two pre-studies covered gamification, citizen perception of the built environment, the use of geolocation technology and location tracking, and also project management relating to budgeting and prioritizing. Based on these takeaways, reviews of both academic and grey literature were carried out, resulting in a set of specifications for the development of an outdoor MAR tool.
The main categories of the specifications were as follows (Imottesjo & Kain, 2018):
1. The tool will simulate built structures and incremental development processes through multi-stakeholder inclusion;
2. The tool will provide immersivity in urban conditions of here and now so that the citizens’
perception of on-site context is not skewed;
3. The tool will have the capacity for on-site AR projection; and
4. The tool will simulate a rule-based process through gaming mechanisms.
Furthermore, twenty sub-categories were developed to specify the detailed requirements needed to fulfill the abovementioned four main categories.
Specification 1.1 is related to how the built structures should be simulated with regards to urban compactness
1. Urban compactness as density and diversity of urban grains (buildings, lots) 2. Urban compactness as diversity and proximity of urban functions and actors Specification 1.2 lists the requirements for the simulation of planning processes
1. Incremental development processes 2. Multi-stakeholder inclusion
3. Rule-based and context-based urban development processes Specification 2 identifies the conditions of immersivity for on-site perception
1. Facilitate perception of urban space, qualities and context, such as urban noise, pedestrian flows, climatic information
Specification 3 details the technical specifications for the on-site AR projection
1. Use mobile devices (smartphones/tablets) as the collaborative interface, with a simple and intuitive user interface for both visualization and design input 2. Use a centralized server and cloud-computing to continuously upload and update design changes done by the multiple stakeholders, to support collaborative design fully
3. Facilitate observation and interaction from different angles (front, back, sides, around corners, etc.) and distances in relation to the real environment for full immersion
4. Use fiducial markers, rather than GPS or building contour tracking, for accurate localization of AR objects. The virtual objects must be perceived in
correct locations and scale, without glitches and lags, when users move around Specification 4 was developed to simulate rule-based urban development processes by including a gaming mechanism in the tool
1. Multi-player role-playing mechanism 2. AR representation from street level 3. Turn-based mechanism
4. Building blocks as economy input, e.g., prices on building blocks 5. Crowd-sourced outcome
6. Site-specific game rules 7. Workshop-based
8. Building blocks representing relevant urban functions for the site 9. Real-world challenges, e.g., ongoing projects, planning rules 10. Create urban emergence
These specifications were then grouped into four design criteria guiding the practical tool implementation: tracking, design elements, UX-I and gaming mechanisms, and data storage and retrieval (see Table 1 in Chapter 4).
5.3.3.2 The first six prototypes: Implementation and assessment of design criteria
As a result of the research-through-design process, the iterative prototyping produced eight prototypes of the Urban CoBuilder based on implementing the design criteria. Of these, the first six were tentative but gradually refined prototypes, and the last two were more developed prototypes, i.e., semi-pilot systems.
The assessment of the first six prototypes regarding the tracking methods indicated insufficient recognizability when using existing built information, such as façades of buildings, as reliable markers in an urban neigborhood scale outdoors. The use of GPS for tracking was too inaccurate for the players to
116
perceive the location of the placed objects in the built environmentnt. Bitonal frame markers were the most reliable and accurate to be usable for the MAR tool, even though the size of the marker mattered for designing in the distance.
When it comes to design elements, the prototype tests highlighted the importance of perception of the scale of the virtual blocks when placed in the environment.
The use of color-coded building blocks was not sufficient and was too abstract for users to understand the built environment.
For UX-I and gaming mechanisms, the prototypes tested various interfaces for the users to add and remove building blocks as well as intuitively select a building block that the user wanted to build with. The display of the remaining budget and building blocks was also tested. It became clear that the main issue that still needed to be dealt with included methods to locate and place the virtual objects on the screen in a more intuitive manner. Furthermore, the use of the grid plane could illustrate the perspective of scale to the ground level. However, when the building block was not on the ground, the location of a hovering block was impossible to perceive. However, this issue was neglected, with the motivation that it is not necessary to be able to simulate a hovering building block, since this would be physically impossible in real-world situations. The addition of stakeholder roles, and projects that belong to these roles with linked different economic rules, was also tested (see Table 4). Here, it was concluded that the many functions and choices that could be made using the tool rendered the screen instead filled with buttons and that the interface needed to be simplified and streamlined through the design of a better user process leading to the selection and changes of different functions.
Assignment Stakeholder Roles Goals Economy/money units available for
1 Municipality Improve green
4 Municipality Increase Housing 1000
Green 50 20
10 Municipality Public facilities 2000
Green 50 20
12 Municipality Increase Housing 1000
Green 50 20
Public
facilities 10 50
Housing 50 20
13 Municipality Public facilities 2000
Green 50 20
Assignment Stakeholder Roles Goals Economy/money
1 Municipality Improve green
4 Municipality Increase Housing 1000
Green 50 20
10 Municipality Public facilities 2000
Green 50 20
12 Municipality Increase Housing 1000
Green 50 20
Public
facilities 10 50
Housing 50 20
13 Municipality Public facilities 2000
Green 50 20 Table 4. Assignments of stakeholder role at each turn
5.3.3.3 The Urban CoBuilder 1:1 scale
This section describes the implemented design criteria for, and the assessment of, the first semi-pilot system: The Urban CoBuilder 1:1 scale MAR tool (see Table
3 in Chapter 4). The second semi-pilot system, the Urban CoBuilder table-top version AR app, is described and assessed in the subsequent section.
For tracking, multiple printed bitonal markers were used in conjunction with a location averaging mechanism, developed through this project, using the gyroscope function of the smartphone. This mechanism enabled the tool to position the location of the user whenever more than one marker was visible on the smartphone’s screen. Through tracking movements supported by the gyroscope function of the phone, the tool could estimate the user location and the position of the camera of the phone when the phone screen lost the sight of any markers, for instance, while looking up to build upwards.
Different types of building blocks were provided with different façade textures to represent the diversity of urban functions, i.e., offices, residential usage, commercial usage, and green spaces (see Figure 44). To provide a basis for the simulation of AR these design elements in the real environment and to aid the player’s perception of scale, an AR grid of three by three meters was overlaid on the site floor. Three by three-meter building blocks could then be placed on this grid (by way of selecting positions indicated by wireframe cubes) and also on top of each other (see Figure 45).
When it came to the UX-I, the possibility to toggle between street view and bird’s eye view was removed to guide the users to explore the virtual object by physically moving around the object and creating a design based on the immediate perception of the object from the street view.
Figure 44. Building functions; office, residential, commercial, and green functions
Figure 45. Street level view
In addition, gaming mechanisms were improved with a system employing a basic form of urban economy to simulate a simple rule, which enabled a player to build according to her/his economic means, depending on the stakeholder role, e.g., a municipal official with public goals linked to housing or services, a private large-scale developer, or a private small-large-scale developer/cooperative (see Table 4). The player got assigned a certain amount of purchasing power and a certain number of playable building blocks representing various urban functions. The price of building blocks varied depending on the role played by the player and the goal of the project; for instance, it was lower to build residential units as a municipal authority with an aim to increase apartments. These basic economic rules would allow the users to prioritize which urban functions should be built within their economic means.
The pilot user testing of the 1:1 scale Urban CoBuilder tool showed the following feedback, including concerns and potential for improvements. The users did not assess the design criterion concerning the data storage and retrieval since it was implemented as the behind the scene base mechanism for saving and uploading the user actions, ex., adding a block, and not as something user could evaluate the functionality of.
Tracking:
Even though the size of the markers was too bulky to be portable, the multiple markers placed on-site, combined with location averaging mechanisms using the gyroscope function of the smartphone, worked sufficiently to stabilize the placed objects locations while the camera leaves the markers.
The integrity of the structure on which the markers were mounted, especially when the marker was placed vertically, was essential. The size of the marker made it easy to become bent during the building sessions, due to its own weight and to weather conditions, such as wind, and this rendered the augmented space to become skewed. Portable smaller standing markers mounted on tripods with three or four faces might make the tracking more stable.
Design elements:
The detailed façade textures on building blocks were assessed as positive, aiding the comprehension of the scale of the augmented building blocks within the built environment. Suggestions were made to include a more extensive choice of façade textures and urban functions, including streets and paths. Even though the scale of the building blocks was received positively, the three by three-meter blocks were too small when working on an urban scale, requiring too many blocks to build a neighborhood scale environment.
UX-I and gaming mechanisms:
In general, the process of starting the app, registering oneself, choosing projects, and understanding the stakeholder roles and the aim of the project was perceived as easy to take in and operate. Also, the mechanisms to place a building block, and remove or change the building block types were all perceived reasonably straightforward.
However, more complexity would be required to include relevant planning rules and economic rules reflecting the real urban planning issues of the test sites. Suggestions were made to add user-generated complexity relating to issues concerning design elements and enabling communication between the players for negotiations and cooperations relating to rules and economy.
Designing a built environment in an urban scale without the possibility for a bird’s-eye view was perceived as unfavorable. The users had desires to view the project from the bird’s-eye view to have a comprehensive picture of what was being built.
5.3.3.4. The Urban CoBuilder table-top version
The Urban CoBuilder, 1:1 scale version, was also modified so that users could test collaborative designing in table-top setting (see Table 3 in Methods chapter).
Regarding tracking, the main modification was the size of the tracking markers for portability, where this version used just one printed bitonal frame marker of 19 x 19 cm.
For the design elements, changes in the building block types and the façade textures of the building blocks were also made, so that the tool would reflect relevant design components for the planning of the study site (see Figure 46). In this case, instead of just including a building’s urban functions as building blocks, façade elements, such as windows, doors, and balconies were developed. These block textures were acquired through analysis of a current building project in the test area, i.e., the BoKlok, IKEA’s approach to residential housing.
The pilot user testing of the Urban CoBuilder table-top version showed that the portability of the tracking markers made it simple to use them during the workshop session, where students were to interview passer-by citizens. Also, scaled-down markers enabled building from top-down (bird’s-eye view) delivered more stable tracking due to smartphones screen remaining in moderately static position, letting the marker to be visible throughout the design sessions.
Figure 46. Façade elements extracted from BoKlok project
For the design elements, the building blocks with façade texture indicating a building element (e.g., door, stairs, windows) were perceived as easy to understand and to build with. It was also found that it was quite useful to build scaled AR models on tables or maps indoors, which could then easily be taken outside for projection on site. However, the smaller scale and bird’s-eye view designing highlighted some issues relating to the scale of the three by three-meter building blocks. Players tended to place the building blocks quicker and to cover large volumes of built areas, and often the smartphone would freeze, unable to handle such a data load.
Regarding the UX-I and gaming mechanisms, the table-top Urban CoBuilder tool was seen as facilitating dialogue with residents passing by for the interviewers on-site, thus turning the AR tool into a MAR tool. However, while the younger generation of interviewed citizens (see Table 5) found the app intuitive and easy to use, it was not comfortable enough for older residents.
AGE Gender M Gender F Reasons for refusal Participated
Under 12 5 2
12-20 3 7
30-40 2
50-60 1
Refused
Mixed 6 Busy
30s-40s 6 Language/suspicious
50s-60s 5 Technical discomfort
Table 5. List of participants and reasons for refusal to use the tool
* The age of the participants were approximated by students. Passerby groups of mixed age, e.g., school children with parents, were categorized as Mixed in the refused section.
Chapter 6.
Discussion
This thesis aimed to contribute to research on processes of incremental bottom-up urban planning and design through citizen inclusion in sbottom-upport of urban resilience based on compact city qualities. The following sections discuss to what extent the objectives of the Ph.D. project (see Chapter 2) have been responded to during the research process and link the results to the state of the art.